Upload
others
View
6
Download
0
Embed Size (px)
Citation preview
Effective Use of Decision Rules in the Hematology Laboratory
Elkin Simson MD, FCAP, FASCPMedical Director
Center for Clinical LaboratoriesThe Mount Sinai Hospital
Vice-Chairman, Dept. of PathologyMount Sinai School of Medicine
New York
Presentation Outline
History of laboratory testingManual, analytical automation, peri-analytical automation, automated information processing
Hematology processes and decisionsHematology Consensus GroupImplementation of consensus rulesMiddleware and LIS
In The Early Days:Manual reagent preparationManual procedures
Reagent & sample pipetting, reagent mixing, precipitation, separation, etc.
Manual calibrants, controlsManual colorimetry
AdjustmentsReadings
Manual calculationsHandwrittenLog tablesSlide rules
Manual Colorimetry (1967)
Remember The Slide Rule?
Improving Accuracy
In The Early Days:Manual colorimetryManual calculations
WrittenLog tablesSlide rules
Manual review for validityIndividual decisionsIndividual performance importantWeariness factorInter-individual inconsistency
The Last Four Decades:Analytical automation
Progressive improvementEnhanced sophisticationUbiquitous in clinical labs
Peri-analytical automationPre- and post-analyticalMore recentMainly robotics
Information processingMainly post-analytical (reporting)Recently peri-analytical (validity assessment)Expert systems
Analytical automationSlow improvement - close to a plateau?
Chemistry Productivity for Specimens & Tests
0
20,000
40,000
60,000
80,000
100,000
120,000
Test
s
0
2000
4000
6000
8000
10000
12000
Spec
imen
s
Tests/employee/yr Specs/employee/yr
1965 1970 1980 1997
Automated Core Laboratory
Peri-analytical automationMajor improvement in productivity
Chemistry Productivity for Specimens & Tests
0
20,000
40,000
60,000
80,000
100,000
120,000
140,000
160,000
1965 1970 1980 199720002003
Test
s
0
2000
4000
6000
8000
10000
12000
14000
Spec
imen
s
Tests/employee/yr Specs/employee/yr
Automated Hematology Laboratory
Hematology Productivity
19,614
30,45334,791 37,966
40,377
26,493
10,000
15,000
20,000
25,000
30,000
35,000
40,000
45,000
1998 1999 2000 2001 2002 2003
Test
s pe
r em
ploy
ee
Peri-Analytical Automation
Hematology Tests & Total Staff
1,253
1,166
1,059
1,279
1,252
1,33262
5248
3834 35
800
900
1,000
1,100
1,200
1,300
1,400
1998 1999 2000 2001 2002 2003
Test
s /y
ear x
100
0
20253035404550556065
Tota
l FTE
's a
t yea
r sta
rt
Tests x1000 FTEs
Hematology Cost Per Test
$2.46
$2.10$1.77
$1.55$1.38 $1.27
$2.02
$3.19
$2.72
$2.30 $2.25 $2.15
$1.00
$1.50
$2.00
$2.50
$3.00
$3.50
1998 1999 2000 2001 2002 2003
Salary/test Total Cost/test
What’s The Next Big Opportunity?
Automated decision rulesAutoverificationAutomated reflex testing
Applicable to labs with LISApplicable to some labs with automated analyzers but not a good LISNext big thing for labs with peri-analytical automationQuality, service and cost benefits
Hematology Process Summary10.8 FTE’s for 3 shifts, 5 day weekTotal worked hours: 403 hours/week
Supervision, training, scheduling: 36 hours (9%)Operations total 148 (37%)
Reagents & maintenance: 54 hours (13%)Analyzers: 47 (12%)Robotics: 16 (4%)Specimens: 31 (8%)
Manual analysis & smear prep: 72 hours (18%)Total Review: 147 hours (37%)
Result review: 72 hours (18%)Smear review: 75 hours (19%)
Note: Prior to autoverification
Results of Application of Mount Sinai Autoverification & Review Criteria
in Hematology
Autoverify if all analyzer results normal6.25% were autoverified
Smear review for analyzer flags and also other criteria
27% required smear reviewManual WBC differential if indicated
8.4% required manual differentials
Validation (Verification) Process
Validate the test resultReview QC results for the runReview analyzer performancePerform delta check
New result may be similar to previously known abnormal resultsSpecimen drawn from wrong patient or sample integrity problems (eg. IV line)
Compare related test resultsA multitude of variables
ActionsActions
Validate (verify) the result and reportRepeat the analysisDilute and re-run the samplePerform reflex test
For most labs, (chem, endo, immuno, etc)
actions involve the same analyzerreflex testing is minimal (e.g. HIV)
Hematology is more complex
What Makes Hematology Special?What to do after automated CBC and differential results have been produced
Verify and reportRepeat with or without dilution on the same or other analyzerExamine a smearVisual differential countManual platelet countOther follow up tests
Validation and Microscopy
Principles for ReviewReduce the number of samples requiring review to the greatest possible extentAnalyzers are fast, with low cost per testHumans are smart but slow, with high labor cost; and good ones are becoming harder to find!Analyzer false positives mainly cause more work; false negatives can harm the patientCriteria for review
What’s safe?What’s most effective?How do we know?Is one lab’s experience enough?
The InspirationThe Inspiration
MAP OF WORLD
WITH LAB LOCATIONS
HIGHLIGHTED
HEMATOLOGYHEMATOLOGYCONSENSUSCONSENSUS
GROUPGROUP
20022002
Mission
The mission of the international consensus group was to develop a set of functional guidelines that would aid laboratories in reducing the number of hematology result reviews and subsequent procedures they are required to perform, without losing any useful clinical data
Participants - 2002 Consensus MeetingIndian Wells, California
Hospital Ramon y Cajal, Madrid SpainLiverpool Hospital, Sydney, AustraliaChinock Health Region, Lethbridge, Alberta, Canada University College Hospital, London, UKMt. Sinai Hospital, Toronto, CanadaAuscorp, Melbourne, AustraliaOregon Medical Laboratories, Eugene, Oregon, USAChildrens Hospital,Denver, Colorado, USA
Grant Medical Laboratory, Columbus, Ohio, USAMayo Clinic, Rochester, Minnesota, USAAlliance Laboratories, Cincinnati, Ohio, USAMt. Sinai, New York City, New York, USADynacare Seattle, Washington USAUC Davis Medical Center, Sacramento, California, USAACL, Milwaukee, Wisconsin, USABarnes-Jewish Hospital. St. Louis, Missouri, USA
2 DAYS IN the HOT DESERT SUN2 DAYS IN the HOT DESERT SUN
20 HEMATOLOGISTS20 HEMATOLOGISTS
200 SODAS TEAS & COFFEE
2000 LAB RULES 2000 LAB RULES (actually ~ 950)(actually ~ 950)
83 rules we could testin our labs
Consensus Study ProtocolFor each lab:1000 patient specimens from daily workload over several daysApprox 200 specimens which had previous results to test delta rulesRecorded all analyzer results, including flags, vote-outs, etc.Reviewed smears from all patients to validate rule performanceCompared each lab’s rules with consensus rulesCompared efficacy of individual consensus rules
Hematology Analyzers in the Study
Advia 120 – Bayer
CellDyn 4000 – Abbott
Gen*S, LH 750 – BeckmanCoulter
SE-9000, XE-2100 - Sysmex
Definition of Positive Smear Findings: Morphology
RBC morphology: 2+/moderate or greaterPlatelet morphology: 2+/moderate or greaterPlatelet clumps: > rare or occasionalToxic granulation: 2+/moderate or greaterDohle bodies: 2+/moderate or greaterVacuoles: 2+/moderate or greaterAny malarial parasite
Definition of Positive Smear Findings:Abnormal Cell Types
Blast ≥ 1Metamyelocytes > 2Myelocytes/Promyelocytes ≥ 1Atypical lymphocytes > 5NRBC ≥ 1 Plasma cells ≥ 1 or more
Rules Summary
Rules separated into five main categoriesWBC parametersPlatelet parametersRBC/Hemoglobin parametersDifferential parameters (#)Suspect codes/flags
Neonates1st time samplesSubsequent sample ( < 72 hrs. )Total of 41 consensus rules
Some Consensus Rules - WBC
WBC >4 and <11 with WBC flags: slide reviewWBC >11 and <50 with no WBC flags: verifyWBC >11 and <50 with WBC flags: slide reviewWBC > 4 <11, first time, no WBC flags: verifyWBC <4, first time, no flags: slide reviewWBC any, incomplete diff: slide review + countWBC < 4, pass delta, same flag except blasts:
verify
Some Consensus Rules - Differential
Neut # < 1.0, first time = slide reviewNeut # < 1.0, pass delta, no blast flag = verifyMono # > 1.5, first time or fail delta = slide reviewMono # > 1.5, pass delta = verifyFlags : Immature granulocytes, or NRBC, or blasts, or atypical lymphs: slide review
Some Consensus Rules - Platelets
PLT <100 and first time: slide reviewPLT >100 and <1000 with no PLT flags: verifyPLT any count and PLT clump flag: slide reviewPLT >100 and other PLT flags, first time analysed: slide reviewPLT <100 and other PLT flags: slide reviewPLT <10 any time: lab SOP, alternate count
Some Consensus Rules - RBC
HGB < 7, first time = slide reviewHGB > 7, no RBC flags = verifyMCV > 105, adult, first time, sample < 24 hrs old = slide reviewMCV < 75, pass delta = verifyRDW > 22, first time = slide reviewRBC fragment flag = slide reviewRBC dimorphic flag = slide review
# SLIDESTOTAL TP FP TN FN REVIEWED
Number 13298 1483 2476 8953 386 3959% 11.2 18.6 67.3 2.9
Truth Tables for All LaboratoriesTruth Tables for All Laboratories
0.0 5.0 10.0 15.0 20.0 25.0
%
Meta
Plt Clump
NRBC
Aniso
Lymph
Myelo
Roulx
Hypo
Poik
Blast
Plasma
Target
Micro
Stomat
Dohle
HJ Body
Pelger
Shisto
Sphero
% FN % Total
Consensus Rules Consensus Rules -- False Neg. RateFalse Neg. Rate
www.islh.org/
Journal Publication of Consensus Rules
The International Consensus Group for Hematology Review: Suggested Criteria for Action Following Automated CBC and WBC Differential Analysis
P. W. BARNES, S. L. MCFADDEN, S. J. MACHIN,
E. SIMSON
Lab Hematol. 2005;11:83-90
How To Implement Consensus Rules?
Reality Bytes
Fewer Medical TechnologistsFewer Medical Technology SchoolsAverage age of an MT is 50More generalists than specialistsInstruments are much more automated than at any other timeMore tests per tech than at any previous time.
A Typical Day
Reporting 200-400 patient profiles per shiftWhile…..
The phones are ringing, the boss wants more – now, residents walk in to look at a smear…
With all of the high tech equipment in most labs today - What are the technologists spending most of their time on?
Data Review
Reviewing each result in light of RULESNormal RangesCritical Values – IP and OPDelta ChecksRelationship of parameters (eg.RBC,HGB,HCT)Dilution Needs or Manual processesLaboratory Review PoliciesPathology Review PoliciesRules for specific units or physicians
How does everyone know the rules?
The Procedure Manuals!
Reality check:Posted sheets on instruments,
cabinets, counter tops, inside lab coats, memory…….
Automated Hematology Analyzers
My hematology analyzer is faster than your hematology analyzer……
Does it really matter that it can do over 100 samples per hour if your brain can only go 35?
So - What’s the answer??
Automation of the Review Process
Maximizes instrument capabilities and logic rules while minimizing patient risk
Reduce technologist interventionsReduce turn around time (TAT)Apply rules in a standardized wayConsistent decision making by all staffIncreases the comfort level of generalistsDecreases tech time of asking “what do I do now” to the senior staff, thus decreasing total turn around time.
Software Solutions
LIS
Middleware + LIS
Middleware (Middle-Hemeware)
Software between the analyzer and the LIS.Information from the analyzer is sent to the middlewareMiddleware software contains all the laboratory’s rulesMiddleware applies the rules to each sampleMiddleware sends the automatically validated results to the LIS, holding back the non-validated samples for technologist reviewDisplays data for samples held back
Middle-Hemeware Displays
Histograms and cytogram scatterplotsPrevious resultsInstrument comments“Next Steps” for staffWhich rules were triggered and the data that met that criteria.Viewable at cell counting workstation, differential workstation and at various other locations (eg. manager’s office, hematopathologist’s office)Specimen results are stored with their respective histograms, scatterplots and all comments.
Advantages of Middle-HemewareIncreased focus on abnormal samplesIncreased release rateFlexible system (add tests, comments)One instrument interface to LIS (standardizes interface specs)Easier creation of rulesRules can be more complex -increases the autovalidation rateReview complete data from multiple analyzers at one workstation
Tech Instructions
Standardizes Next Steps for TechsCan be as simple as “scan slide”More complex: R/O cold agglutin or lipemia per lab SOP # 566.Displays on Results screen –easily visible to techs.Useful to all staff from senior to new techs.
HGB Consensus Rule
Consensus Rule #9
HGB is less than 7.0g/dl or > 2g/dl above upper reference range for age and sex.
Tech instructions: Slide Review if first time. Verify sample integrity if indicated.
Rules example #2
Consensus rule #35Atypical/Variant Lymph Flag (suspect
comment)Tech Next Step: Slide Review
Easily handled by Middle-Hemeware
Handled by LIS if it can accept instrument comments and histograms and can display tech instructions
Rules example #3Parameter Combinations
Consensus Rule #36Atypical/Variant Lymph Flag, with
previous confirmed result and positive delta flag for WBC.
Tech instruction: Slide Review
Able to write this rule in Middle-Hemeware, difficult in LIS systems
Rule Example #4
Tailoring the consensus rules to physicians
Consensus rule #4: WBC <4.0 or > 30.0 and first time
Tech instruction: Slide Review*********************************WBC <4.0 or > 20.0 and Dr. JonesTech instruction : Slide Review
Summary: Realized Benefits of Middle-Hemeware
Able to view multiple hematology analyzers from one workstation, with no paper printoutsEasier creation of rules than most LIS systems – can specify rules for suspect flags, as well as combination of parameters.Standardized rules for review
Summary:Realized Benefits of Middle-Hemeware
Decreased overall turn around timeMore work per technologist without stress & errorsWhen reviewing results, all information is at hand, not just the numbers.Increase auto-validation rateTechnologists able to spend more time focusing on abnormal samples
Conclusions
The hematology laboratory is a more complex operation than mostStandardization of decision rules has been initiatedAutomation of rules is availableIn most situations, Middleware provides a preferred solution